Positive selection results obtained using HyPhy from a dataset excluding individuals with sequences enriched for APOBEC hypermutation, as well as from the complete dataset including the hypermutated sequences (the latter sites are indicated with +).

Sites identified using the maximum likelihood phylogeny-based method are indicated with ∼ . The location, timing, and mutational patterns observed are provided as well as an indication of whether CTL testing was carried out or not.
Key for Table 2
HXB2: Coordinates listed according to HXB2 numbering (http://www.hiv.lanl.gov/content/sequence/LOCATE/locate.html)
Data: Coordinate listed according to the protein alignment of all 81 acutely infected subjects.
Location: Region in Envelope.
Posterior Probability: Sites identified in HyPhy with a posterior probability>0.5 are included in this table.
P-value/q-value: The p and q values from the Phylogenetic analysis for a specific mutational pattern being associated with an earlier or later Fiebig stage.
Number of subjects with variation: Out of the 81 subjects, the number that had any variation in this site.
APOBEC3 in subject's mutations: The number of subjects among those that vary in a given position that have a G to A change in the context of an APOBEC3 motif.
Timing: The selection results were obtained using all samples in Fiebig stages I–V.
For the phylogenetic method a change at the site was enriched in the range of Fiebig stages shown.
Number of subjects with a mutational pattern: This summarizes all individuals with changes found in this position in the data. Italics means a change was found more than once in at least one person with the pattern.
Posterior probabilities: The number of individuals that have a change from the most common amino acid to another. If there was more than 1 change in any of the individuals, it is noted in italics.
 Example 1: Site 651: 2 N to S, 1 S to G means: Two people had N as the most common amino acid, but S was present. PRB931 had 17 N and 2 S, PRB956 had 25 N and 1 S. Because there was more than one S in one of them, it is in italics. One person, 700010077, in noted as S to G, and had 51 S and 1 G.
 Example 2: Site 518: 2 M to I, 7 M to V means: Nine people had M as the most common amino acid, and each of the nine had a single variant among their sequences, and it was I in 2 of them, or V in nine of them. For example, subject 1012 had 42 sequences, with 41 M and 1 V.
Phylogenetic method: The within-subject change that was observed to be enriched either early or late Fiebig stages.
 Example: Site 460: 2 Away from N means: In the subjects that had changes in this position, the ancestral state of the founding virus (the transmitted virus) was most likely to be N, and in both cases at least one change away from N was observed. Because it is italics, at least one of the two carried more than one mutation from N in this position.
CTL testing: If a variant was found multiple times in a patient or embedded in the context of additional proximal changes, peptides spanning the region were generated and T-cell assays were perform (see Table 3).
*In Keele et al., we noted patients bearing these changes might have been infected with more than one closely related form. Alternatively, early selection or maintenance of a very early mutational event might be giving rise to the pattern.
**This site was difficult to align in a few patients due to a frameshifting insertion of an A in a string of As in the primary sequences, thus some of 7 noted E to K changes were actually due to a frameshifting indel.
***In an additional subject there was an ambiguous base call in this position.